Skip to main content

Applications per Credential

Average number of distinct applications one credential reaches.
Warning

No longer relevant. Use Multi-Service Users instead. This metric estimates breadth from summary tables that have no user IDs, and internal CanadaLogin components and multi-application services inflate it. The IBM Verify events pipeline now lets us count services per user directly.

The average number of distinct applications a single credential reaches.

This metric measures how far the typical user spreads across CanadaLogin-integrated applications. It works because per-application unique_users counts each person once per application, while the CanadaLogin-wide count counts each person once overall: a user active on three applications adds three to the per-application sum but only one to the overall total. Dividing the two recovers the average number of applications per credential.

A value of 1.0 means the typical credential touches exactly one application; 2.0 means two, and so on.

Calculation

Sum the latest year-to-date unique users across every application (from events_app_login_counts), then divide by the latest CanadaLogin-wide year-to-date unique users (from events_auth_total_logins). Both inputs are cumulative ytd_unique_users fields, so take the latest value per source rather than summing across days.

WITH app_latest AS (
    SELECT
        application_name,
        ytd_unique_users,
        ROW_NUMBER() OVER (
            PARTITION BY application_name
            ORDER BY from_date DESC
        ) AS rn
    FROM ibm_verify_events_transformed.events_app_login_counts
),

app_total AS (
    SELECT SUM(ytd_unique_users) AS sum_app_unique_users
    FROM app_latest
    WHERE rn = 1
),

overall_total AS (
    SELECT ytd_unique_users AS overall_unique_users
    FROM ibm_verify_events_transformed.events_auth_total_logins
    ORDER BY from_date DESC
    LIMIT 1
)

SELECT
    CAST(app_total.sum_app_unique_users AS DOUBLE)
        / NULLIF(overall_total.overall_unique_users, 0) AS applications_per_credential
FROM app_total
CROSS JOIN overall_total;
sum_app_unique_users <- tbl(con, in_schema("ibm_verify_events_transformed", "events_app_login_counts")) |>
  group_by(application_name) |>
  slice_max(
    from_date,
    n = 1,
    with_ties = FALSE
  ) |>
  ungroup() |>
  summarise(
    sum_app_unique_users = sum(as.double(ytd_unique_users), na.rm = TRUE)
  ) |>
  pull(sum_app_unique_users)

overall_unique_users <- tbl(con, in_schema("ibm_verify_events_transformed", "events_auth_total_logins")) |>
  slice_max(
    from_date,
    n = 1,
    with_ties = FALSE
  ) |>
  pull(ytd_unique_users)

sum_app_unique_users / overall_unique_users

Interpretation notes

This is a population average, not a per-user count: a small group of power users spanning many applications can pull it up while most credentials stay on one.

The figure is inflated by internal CanadaLogin components, especially the migration app (GCS Migration Solution) and the profile management app (GC Sign In - Profile Management App). A user who reaches a single public service is still recorded as a unique user in the migration app when their credential is migrated from ECM, and in the profile management app when they manage their account, so they count two or three times in the per-application sum while counting once overall. Related applications of one service, such as MyCGC’s two, also count separately. For a clean measure of breadth across public services, use Multi-Service Users, which counts services per user directly.

Swap ytd_unique_users for mtd_unique_users to measure the same breadth over a different window.

Alternative names

  • average applications per user
  • application breadth per credential
  • multi-app usage
  • cross-application reach
  • apps per account